INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The gram-negative obligate anaerobe Porphyromonas gingivalis is one of the organisms most strongly associated with chronic adult periodontitis (59). P. gingivalis expresses numerous potential virulence factors, such as fimbriae, hemagglutinins, lipopolysaccharides, and various bacterial enzymes and proteases which are capable of hydrolyzing collagen, immunoglobulins, iron-binding proteins, and complement factors (2). One enzyme that may contribute to the virulence of this organism is superoxide dismutase (SOD). This enzyme, along with catalase and peroxidase, belongs to a specific cellular system that has evolved for cellular protection against oxidative stress. Interestingly, a previous study was unable to detect either catalase or peroxidase activity in P. gingivalis (2).

SOD has been found in nearly all aerobic organisms studied, as well as in numerous obligate anaerobes that can tolerate transient exposure to oxygen (11, 28). It acts to scavenge molecular oxygen and its univalent reductants, thereby protecting the cell from the harmful effects of these reactive oxygen species (ROS). The absence of SOD activity causes a variety of oxygen-dependent phenotypic defects in Escherichia coli, including a high rate of spontaneous mutagenesis, severe defects in amino acid biosynthesis, and structural instability of the cell envelope (27). Interestingly, overexpression of the iron-containing SOD (FeSOD) in E. coli leads to increased oxygen sensitivity (54).

Eubacteria and archaea typically contain cytosolic SODs that require manganese or iron as a cofactor (MnSODs and FeSODs, respectively). MnSOD is also found in the mitochondria of eukaryotes. MnSODs and FeSODs have very similar structures and presumably evolved from a common ancestral gene, and both may be present in an organism, such as those encoded by the sodA and sodB genes in E. coli (28). The conserved and well-distributed nature of sodA, sodB, and other genes involved in the oxidative stress response suggests the importance of these genes to the organisms. For instance, although Legionella pneumophila expresses a periplasmic copper- or zinc-containing SOD (CuZnSOD) (encoded by the sodC gene), the cytoplasmic FeSOD is essential for viability under normal culture conditions (51). Originally considered to be specific to eukaryotes, the sodC gene has been identified in a dozen or so bacteria to date (28). In P. gingivalis, the cytosolic SOD is encoded by one gene, sod, and utilizes Fe or Mn for activity (3). Two other members of the Bacteroides genus and one representative each from the Propionibacterium, Streptococcus, and Methylomonas genera also share this property (39).

The antimicrobial activities of monocytes and polymorphonuclear phagocytes (PMNs) have been broadly characterized as being either oxygen-dependent or oxygen-independent systems. Oxygen-dependent systems include the production of ROS, mediated by the phagocyte oxidative burst, and the reactive nitrogen intermediate, nitric oxide (49). ROS include superoxide (O₂·), hydrogen peroxide (H₂O₂), singlet oxygen ('O₂), and hydroxyl radical ( ^· OH). Numerous microorganisms counter these antimicrobial species by impairment of phagocytosis, alterations of microbicidal activity, or attenuation of the oxidative burst. Specifically, P. gingivalis proteolytic activity inhibits the generation of ROS in PMNs, thus suppressing bactericidal ability (44). In addition, purified SOD from P. gingivalis added exogenously partially protected the organism from this method of killing by PMNs (4).

Another product of activated phagocytes, nitric oxide (NO), has been shown to be cytostatic or cytotoxic for a variety of invading microorganisms. Its proposed actions include the interference with iron-dependent enzyme functions in bacteria and reaction with superoxide to form peroxynitrite, which in turn decomposes to form the highly reactive hydroxyl radical and nitrogen dioxide (22). The biological significance of these reactions in humans remains to be determined; however, the important role of this system in the mouse model is clear (49).

In this study, we describe the utilization of an isogenic SOD-deficient mutant (CAL1) and a reporter gene fusion construct (CAL2) to determine the role of SOD in the periodontopathogen P. gingivalis. We have confirmed previous findings that SOD is essential for aerotolerance (43) and now report that the P. gingivalis enzyme appears to confer different levels of protection against various ROS. Furthermore, in vitro studies of this organism suggest that SOD may not be essential for resistance to killing by neutrophils, supporting similar findings for other organisms (45).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. E. coli strains were maintained in Luria-Bertani broth and on Luria-Bertani agar plates at 37°C with the appropriate antibiotics. P. gingivalis 381 was grown at 35°C anaerobically in a Bioblend mixture (5% carbon dioxide, 10% hydrogen, and 85% nitrogen) (Cryogenic Supply, Buffalo, N.Y.) on enriched tryptic soy agar (ETSA) plates (Difco Laboratories, Detroit, Mich.) as previously described (60). Broth cultures of P. gingivalis were maintained in enriched tryptic soy broth (ETSB) under the same conditions as the plates. All chemicals were obtained from Sigma (St. Louis, Mo.) unless stated otherwise.

Construction of an isogenic SOD-deficient mutant (CAL1). Plasmid pCC19, containing the sod gene, was constructed in our laboratory and described previously (11). A 2.3-kb HindIII-PstI fragment containing the intact sod gene and its flanking regions from P. gingivalis ATCC 53977 was initially inserted into pUC19. Plasmid pVA2198, containing an ermF-ermAM cassette, was obtained from F. Macrina (Medical College of Virginia, Richmond) and has been described previously (19). pCC19 and pVA2198 were maintained in E. coli JM109 in the presence of 50 µg of ampicillin per ml and 300 µg of erythromycin per ml, respectively. The EcoRI-BamHI fragment containing the ermF-ermAM cassette was isolated from pVA2198 and inserted into the EcoRI-BglII-cleaved sod gene in pCC19, thereby displacing an internal fragment of the gene. This plasmid, pCC19Em, was linearized and then used to electroporate P. gingivalis 381 essentially as previously described (19). Strain 381 was used for mutant construction, since this strain is more highly transformable than strain 53977. The transformants were spread onto ETSA plates containing erythromycin (5 µg/ml) and gentamicin (25 µg/ml), and the plates were incubated anaerobically for 10 to 14 days. Antibiotic-resistant colonies were selected for Southern blot and functional analyses. Loss of SOD activity was confirmed by two different methods. Crude cell extracts were prepared following French press disruption in the presence of protease inhibitors (1 mM leupeptin or protease inhibitor cocktail). A xanthine-xanthine oxidase coupled spectrophotometric assay (38) and a nondenaturing polyacrylamide activity gel (8) were employed to measure SOD activities. For the latter, the stacking gel (4.8% polyacrylamide) was adjusted to pH 8.3, and the resolving gel (10% polyacrylamide) was adjusted to pH 9.0.

Construction of a P. gingivalis sod::lacZ strain. The translational fusion vector pMC1871 (55) (Pharmacia Biotech, Piscataway, N.J.) was digested with BamHI to yield a fragment containing the promotorless lacZ gene. This fragment was inserted into the BglII site within the sod gene to yield an in-frame fusion construct, pCC19Z. The shuttle vector pKDCMZ (courtesy of K. Nakayama, Kyushu University, Fukuoka, Japan) (43) was digested with PstI and SmaI to allow for insertion of a PstI-MscI fragment from pCC19Z containing the sod::lacZ fragment, to yield pKDSODZ. pKDSODZ was utilized as a suicide vector along with the mobilizing plasmid R751 (40) for conjugal transfer of the sod::lacZ fusion construct. This was accomplished by mixing 2 ml of a mid-log-phase culture of P. gingivalis 381 with 2 ml of a mid-log-phase culture of E. coli DH5 (containing pKDSODZ and R751) and with chloramphenicol (25 µg/ml) and tetracycline (10 µg/ml). The sample was centrifuged for 5 min at 3,000 × g at 4°C, resuspended in 0.2 ml of ETSB, and plated onto ETSA plates. The plates were incubated for 2 h aerobically at 37°C and then transferred to an anaerobic chamber for 36 h. Bacterial growth on plates was harvested with a sterile cotton swab and resuspended in 1 ml of ETSB, and 0.2-ml aliquots were spread onto ETSA plates containing gentamicin (100 µg/ml) and erythromycin (5 µg/ml) and grown anaerobically for 10 to 14 days at 35°C. Colonies were screened by Southern blot analysis for preliminary confirmation of the correct construction.

Southern blot analysis. P. gingivalis chromosomal DNA was isolated from late-log-phase broth cultures by using the recommended protocol supplied with the Puregene DNA Isolation Kit (Gentra, Minneapolis, Minn.). After digestion with selected restriction enzymes, DNA was loaded at 2.5 µg/lane onto 0.7% agarose gels, and electrophoresis was performed. Gel preparation, transfer onto Hybond-N⁺ membranes (Amersham, Arlington Heights, Ill.), probe labeling, hybridization, and detection with the enhanced chemiluminescence (ECL) system were performed as recommended by the supplier (Amersham).

-Galactosidase assays. The -galactosidase assay was a modification of a previously published protocol (41). Cultures of P. gingivalis were harvested (1.5 ml) after growth under the indicated conditions, the optical density at 600 nm (OD₆₀₀) was recorded, and the culture was centrifuged for 8 min at 16,000 × g at 4°C. The pellet was resuspended in 0.9 ml of buffer Z (41) and 0.1 ml of toluene, vortexed vigorously, and placed in a shaking incubator at 30°C for 45 min, with vortexing every 15 min. Samples were then centrifuged for 8 min at 16,000 × g at 4°C, 250 µl was aspirated (including the top organic phase), and 250 µl of ONPG (o-nitrophenyl--D-galactoside) was added and vortexed vigorously. Samples were then placed at 30°C in a shaking incubator for 15 to 45 min, and then the reaction was stopped with 0.5 ml of 1 M Na₂CO₃ and the sample was centrifuged for 8 min at 16,000 × g at 4°C (this obviated the need for spectrophotometric measurement at 550 nm). One milliliter was used for OD₄₂₀ measurement, and calculations were performed by using established formulae; enzyme activities were expressed in Miller units (41). All assays were performed in duplicate or triplicate and repeated at least three times.

Comparison of environmental stresses on sod expression. Growth of P. gingivalis cultures was determined by turbidimetric measurements of OD₆₀₀; viability was measured by plating serial dilutions of cultures onto ETSA plates and incubating them anaerobically for 10 to 14 days for viable-cell counting. All of the experiments were conducted by adding the indicated compounds to cultures in ETSB medium. Aeration was achieved by gently vortexing broth cultures in 50-ml tubes and placing them in a shaking incubator at 200 to 250 rpm. Various concentrations of hemin (0 to 100 µg/ml) were tested, with and without the addition of the ferrous iron chelator 2,2'-dipyridyl (DPD) (125 µM). Iron limitation was achieved by successive passaging of cultures in ETSB containing 10 µg of hemin per ml and 125 µM DPD by previously described methods (6). Hydrogen peroxide (30%, wt/wt) was added to achieve the indicated concentrations, and pyrogallol was used to generate extracellular superoxide (36). All samples were incubated in duplicate, and the experiments were repeated at least three times.

Human PMN bactericidal assay. The PMN bactericidal assay is a modification of an established protocol (62). Thirty milliliters of peripheral venous blood was obtained, after informed consent was obtained from the single volunteer, using a 19-gauge needle to avoid activation of neutrophils. Three milliliters of 1.5% Na₂EDTA-phosphate-buffered saline (PBS) was added to the blood sample and centrifuged for 20 min at 1,000 × g at room temperature, and the top two phases were discarded. Forty milliliters of 2% gelatin in a 0.9% NaCl solution was added, gently mixed, and then placed in a 37°C water bath for 45 min. The top phase was collected and centrifuged for 10 min at 1,000 × g at room temperature. The pellet was collected and resuspended in 9 ml of cold distilled water for 30 s, and erythrocyte lysis was stopped by the addition of 3 ml of cold 3.5% NaCl together with 0.5 ml of incomplete PBS. The sample was centrifuged for 5 min at 1,000 × g at room temperature, the supernatant was aspirated, and the pellet was resuspended with complete PBS to a concentration of 10⁷ cells per ml. Wild-type or CAL1 (50 µl; 2 × 10⁷ CFU) samples were added to mixtures containing combinations of human agammaglobulinemic serum (5%, as a complement source; from M. E. Wilson, UMDNJ-NJDS, Newark, N.J.), rat anti-P. gingivalis 2561 (0 to 100 µg/ml; from R. J. Genco, SUNY at Buffalo, Buffalo, N.Y.), and 10⁷ neutrophils in complete PBS. The contents of each tube were mixed thoroughly following bacterial addition, after which (time zero) 50 µl was rapidly withdrawn and transferred to tubes containing 4.95 ml of sterile H₂O with 0.2% filter-sterilized bovine serum albumin. Following 1 min of incubation at room temperature with gentle mixing, 50 µl of the PMN lysate was removed and transferred to sterile tubes containing 1 ml of prereduced ETSB. Aliquots of 0.1 ml (~250 to 300 CFU) were plated onto ETSA plates and immediately placed in an anaerobic chamber. Following removal of time zero samples, the polypropylene tubes containing samples were placed onto a rotator with continuous end-over-end rotation at 8 rpm in a humidified CO₂ incubator at 37°C. Samples were withdrawn at 20, 40, and 60 min and processed identically to the time zero samples. Plates were incubated anaerobically for 10 to 12 days, and colonies were counted.

Spontaneous mutations rates. Mutagenesis, as measured by development of rifampin resistance, was determined for wild-type 381 and mutant CAL1. Overnight cultures were used to inoculate prereduced medium and allowed to grow to mid-log phase, as determined by OD₆₀₀ values. Cultures were then exposed to shaking aerobic culture conditions at 37°C for 15 min and then placed in a GasPak holding jar containing H₂ and CO₂ (BBL GasPak Plus, Becton Dickinson, Cockeysville, Md.) until the cultures reached mid-log phase. OD₆₀₀ values were recorded, and serial dilutions of respective cultures were performed, with subsequent plating on ETSA plates containing gentamicin (25 µg/ml) for CFU determination. Two hundred microliters of undiluted samples was plated on ETSA plates containing gentamicin (25 µg/ml) and rifampin (1 µg/ml) and placed under anaerobic conditions as described above. Mutation rates were calculated by dividing total rifampin-resistant CFU by total CFU per ml of culture.

RNA isolation and Northern blots. Total RNA was isolated from P. gingivalis cells grown to mid-log phase by a modification of a previous method (60). The MscI-BglII sod fragment was isolated from pCC19, and 300 ng was used to probe the membranes by using the ECL detection system as described previously (Amersham).

Transcription start site determination. An 18-mer oligonucleotide (SOD; 5'-CGTGCCGATGATGAGCTT-3') corresponding to a location approximately 110 nucleotides downstream of the sod start codon was used as a primer. In addition, a 17-mer oligonucleotide (UPSODR; 5'-GCTGTCATCAGTCACGT-3') and a 19-mer oligonucleotide (UPSODR2; 5'-GGCTGTGGTACCTTGAAGA-3') corresponding to positions approximately 100 and 40 nucleotides, respectively, downstream of a putative transcription start site were used as primers and end labeled with [-³²P]ATP (DuPont, NEN Research Products, Wilmington, Del.). Reverse transcriptase (Gibco BRL) was used to extend these primers to produce cDNA complementary to P. gingivalis mRNA following annealing. After RNase A treatment, the resulting end-labeled cDNA was electrophoresed on 5 and 8% Long Ranger-urea gels (FMC BioProducts, Rockland, Maine). Dideoxy sequencing reaction mixtures with pCC19 and the above-described primers and with M13mp18 and the 40 primer were also electrophoresed for reference. Dried gels were exposed to Kodak X-Omat AR film, and the sod transcription start site was determined by counting the bases from the respective primers as described previously (56).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of the SOD-deficient mutant (CAL1). The sod gene from P. gingivalis 53977 was previously isolated and shown to be oriented between the prtT gene and an uncharacterized open reading frame (ORF) (Fig. 1) (11). Insertional inactivation of the strain 381 sod gene with an Em^r cassette was carried out as indicated in Fig. 2. We confirmed a double-crossover recombination event by Southern blot hybridization with DNA probes derived from fragments containing the sod gene of P. gingivalis 53977 and the Em^r cassette (data not shown). This event resulted in the deletion of an internal fragment of the sod gene and the insertion of the Em^r cassette into the chromosome of P. gingivalis 381, producing a stable, insertionally inactivated mutant (CAL1) resistant to erythromycin (up to 25 µg/ml) (Fig. 1 and 2). In addition, the loss of SOD activity was confirmed by using both a SOD activity gel and a spectrophotometric method (data not shown). As has been previously observed for this organism (43), deletion of the sod gene resulted in a profound loss of aerotolerance. In the present study, a loss of viability of more than 4 orders of magnitude within 1 h of aerobiosis was noted. Apart from phenotypic changes that could be directly ascribed to loss of SOD function, CAL1 appeared to maintain properties essentially similar to those of the wild type. Differences were often seen in the form of a prolonged lag phase, difficulty in recovery from stationary-phase cultures, and decreased maximal OD in stationary phase.

View larger version (5K): [in this window] [in a new window]   FIG. 1.   Schematic of the P. gingivalis 53977 chromosome region containing the sod gene.

View larger version (22K): [in this window] [in a new window]   FIG. 2.   Plasmid maps for construction of the sod mutant (CAL1).

Role of SOD in sensitivity to environmental stress. In virtually all organisms studied to date, SOD expression has been to shown to be affected by specific environmental conditions as well as the growth rate of the organisms. Therefore, to test the potential inability of CAL1 to tolerate selected environmental stress, we compared the growth and viability of the wild type and CAL1 in the presence of elevated temperatures, ROS, and human PMNs.

Construction and characterization of a sod::lacZ variant. In order to examine the regulation of sod expression, a sod::lacZ translational fusion strain was constructed (Fig. 3). A single crossover event resulting in a P. gingivalis sod::lacZ genotype was confirmed by Southern blot hybridization with DNA probes derived from the sod and lacZ genes (data not shown). Furthermore, an intact copy of the sod gene was generated in CAL2 (sod::lacZ), as suggested by Southern blot analysis and subsequently confirmed by normal aerotolerance (data not shown). To our knowledge, this was the first reported demonstration of the utilization of the lacZ gene as a reporter gene in P. gingivalis (32). Subsequently, lacZ reporter constructs for other P. gingivalis genes have been reported (29, 64).

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Construction of the sod::lacZ strain (CAL2).

View larger version (15K): [in this window] [in a new window]   FIG. 4.   Effect of aerobiosis and temperature on sod::lacZ expression in CAL2. Mid-log-phase CAL2 cells were grown for 2.5 h under the indicated conditions, harvested, and assayed. -Galactosidase activity is displayed in Miller units, and the values shown are for duplicate samples from a single experiment and are representative of three separate experiments. Positive standard deviations are indicated by the dark boxes.

View larger version (21K): [in this window] [in a new window]   FIG. 5.   Growth-dependent expression of sod::lacZ in CAL2. Prereduced ETSB was inoculated with an overnight CAL2 culture, which was then grown anaerobically at 35°C, harvested at the indicated time points, examined for growth (OD600), and assayed for -galactosidase activity. Activity is displayed in Miller units, and the values shown are for duplicate samples from a single experiment and are representative of three separate experiments. Standard deviations ranged from 0.49 to 3.52.

View larger version (11K): [in this window] [in a new window]   FIG. 6.   Effects of pH on sod::lacZ expression in CAL2. Mid-log-phase CAL2 cells were transferred to buffered ETSB at the indicated pH, grown for 8 to 12 h anaerobically, and harvested at mid-log phase. -Galactosidase activity is displayed in Miller units, and values shown are for duplicate samples from a single experiment and are representative of three separate experiments. Positive standard deviations are indicated by the dark boxes.

sod transcription initiation. Since the regulation of sod expression appears to occur at the transcriptional level (5), it was of interest to attempt to identify potential regulatory regions upstream of the sod gene. Therefore, the transcription start site of the sod gene was determined to be 315 bp upstream of the sod start codon (Fig. 7) and to be within an uncharacterized upstream ORF, ORF2 (Fig. 1). There were no obvious consensus 10 and 35 sequences when the sequence was compared to previously suggested promoter region sequences for other P. gingivalis genes (29). The adenosine 315 bp upstream of the start codon appears to be the major transcription start site for sod, which would led to a transcript of approximately 0.9 kb. This was the size observed for sod mRNA on Northern blot analysis (reference 5 and data not shown). The guanosine immediately downstream of the A base may act as a secondary start site. Reverse transcription-PCR results obtained by using primers present in the putative transcript further substantiated these findings (data not shown). No DNA-binding consensus regions for known SOD regulatory proteins in other organisms could be identified upstream of the sod gene. Specifically, a search for consensus sequences for Fur (58), integration host factor (46), and Sox (63) operators revealed no significant homology.

View larger version (47K): [in this window] [in a new window]   FIG. 7.   Primer extension analysis of the sod transcription start site. (A) Region upstream of the sod coding sequence. The location of the major transcription start site (the A located within the region corresponding to the UPSODR2 reverse primer) 315 nucleotides upstream of the sod start codon is shown by an asterisk. UPSOD1 and UPSOD3 forward primers were used for reverse transcription-PCR and served to confirm the primer extension results. (B) Primer extension analysis of the sod transcript. Three oligonucleotide primers were used. Lane 1, primer extension product of the UPSODR reverse primer. The major transcriptional initiation start site is indicated by the upper band. The UPSODR2 primer produced no product, as expected. The SOD primer produced a band consistent with results from the other primer extension reactions. The transcriptional initiation start site was determined by counting the bases of the sequence of M13mp18 from the 40 primer.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Periodontal disease is a multifactorial entity in which complex microbial interactions are profoundly significant. The aerotolerant P. gingivalis appear to be capable of countering and even modifying the host response through a variety of mechanisms (66). While the importance of these potential virulence factors is clearly related to the survival ability of these organisms, it is also essential to consider the conditions that these organisms encounter in their colonization of the host. In particular, transmission and initial colonization events may pose the most crucial oxidative stresses to anaerobic organisms. With this in mind, it was logical to propose that the SOD enzyme may play its most significant role in these early events.

Our studies have demonstrated that while the P. gingivalis SOD is relatively selective in its response to a variety of environmental stresses, it obviously mediates at least one extremely important property, that of aerotolerance. Given the unique ecological demands that the oral cavity poses for anaerobic bacteria such as P. gingivalis, this aerotolerance may be as significant as any of the other potential virulence factors in this organism's armamentarium. This organism must first traverse the hostile oral cavity, an aerobic environment constantly bathed with saliva containing antimicrobial properties. Once overcoming this challenge, the organism must adhere to and colonize a complex biofilm, dental plaque. Although recent studies have detected P. gingivalis in supragingival plaque (65), itself an extremely competitive ecosystem, the organism is considered to be primarily a subgingival resident. Therefore, this organism must compete for a niche in the gingival crevice in the presence of host defense responses as well as in competition with other bacterial species.

SOD expression in P. gingivalis appears to be responsive to increased temperature, a key component of inflammation. This property appears to be distinct from the known heat shock response in this organism (35). Previous studies have suggested that increased temperatures may promote higher reactivity of oxygen radicals (9) and may induce gene expression through the alteration of DNA supercoiling (61).

The hydrogen peroxide biomodal killing pattern observed in E. coli (24) does not appear to be exhibited by P. gingivalis. In addition, whereas E. coli sodA sodB was more sensitive to mode-one killing by hydrogen peroxide (26), P. gingivalis CAL1 was no more sensitive than the parental organism. A pattern more suggestive of a linear response to hydrogen peroxide concentrations was observed in our study for P. gingivalis. Compared to E. coli, this organism was sensitive to lower levels of the peroxide, which may be expected from the obligate anaerobe.

A large and diverse range of microorganisms have evolved to utilize the long-lived mononuclear phagocytes as target cells to fulfill an obligate requirement for an intracellular environment in which to survive and replicate (49). While P. gingivalis certainly does not require an intracellular environment for survival, it is reasonable to consider whether this organism can evade prompt destruction after phagocytosis and possibly even persist intracellularly. During phagocytosis, ROS are confined to the phagocytic vacuole and serve as agents highly toxic to the internalized microbial agent. Some bacteria have developed means to resist the oxidative burst encountered within the phagolysosomes, and it has been suggested that SOD may protect Nocardia asteroides (7, 18) and Listeria monocytogenes and Shigella flexneri (42) from phagocytic killing as well. The SOD-deficient mutant, CAL1, did not appear to be more susceptible to killing by PMNs; this result is similar to that obtained in one study of E. coli deficient in the sodB gene (45). We have also found that extracellular ROS generated by pyrogallol killed the CAL1 mutant at rates similar to those for the parental organism. Therefore, these results indicate that the sod gene may not protect P. gingivalis from extracellular ROS. A previous study of P. gingivalis suggested that SOD is protective against killing by PMNs (4). However, in those experiments, purified P. gingivalis SOD was added to cultures, which allowed for the extracellular quenching of ROS, thus likely creating a nonnatural system for accurate assessment of this enzyme's function. However, it is possible that a periplasmic SOD similar to CuZnSODs, or even possibly nonspecific mechanisms, could play such a role in these organisms. No evidence for such an enzyme in these organisms has been reported, and our attempts to demonstrate such activity were inconclusive. The PMN is essentially a suicide cell, releasing an extraordinary oxidative burst that may override the antioxidant abilities of the cytoplasmic SOD in P. gingivalis, as suggested for other organisms (53). However, it is also possible that the SOD may protect the organism against oxidative killing by other phagocytes such as macrophages.

In addition, recent studies have demonstrated this organism's ability to invade, persist, and even replicate within certain host cell lines in vitro (31, 52). In a gingival epithelial cell culture, this organism was capable of intracellular replication and persisted within these cells over an 8-day period. It was found in the cytoplasm and was not confined by a membranous vacuole (34). It is reasonable to speculate that SOD may play a role in such intracellular survival.

Protection against oxygen-dependent DNA damage appears to be an important property of SOD in many organisms (25, 26). Indeed, a 6.8-fold-higher rate of spontaneous mutation to rifampicin resistance was observed in the sod mutant. P. gingivalis appears to have only a single sod gene, compared with the three found in E. coli. Two of these SODs in E. coli are found in the cytoplasm, and previous studies have demonstrated that MnSOD is highly protective against oxygen-dependent DNA damage, whereas the other, FeSOD, appears to confer modest protection at best (relative mutation rates of 8.7 and 0.9, respectively) (16). However, deletion of the two cytoplasmic-SOD genes resulted in a relative mutation rate of 41 compared to that for the wild-type strain. It was suggested that the greater protection provided by MnSOD in the sodB mutant was directly related to its inducibility. The observations that P. gingivalis contains only a single sod gene and that the CAL1 mutant exhibited a relatively high mutation rate compared to the wild type strongly suggest a critical role for SOD in protection against oxygen-dependent DNA damage in this periodontopathogenic bacterium.

In this study, we examined the regulation of expression of the sod gene in P. gingivalis. Although not well characterized in most organisms, the regulation of this gene relative to stress responses has received much attention in E. coli. For example, whereas the activity of FeSOD is similar under almost all growth conditions, the activity of MnSOD is modulated by oxidative environments both transcriptionally and posttranslationally in a metal-dependent fashion (47). Six global effectors of transcription are presently known to affect MnSOD expression (13). Homology searches for other superoxide stress-inducible genes and their corresponding transcriptional regulators in well-studied prokaryotes by using the P. gingivalis genome sequence (TIGR) did not yield any significant homologies. Interestingly, a similar search for genes involved in the hydrogen peroxide stress response, including the transcriptional regulator oxyR, did yield potential homologues, and we are currently investigating these findings. The possibility that the hydrogen peroxide stress response elements are present in this organism remains to be confirmed. However, if this indeed was the case, it would be of considerable interest, since P. gingivalis appears to be much more sensitive to hydrogen peroxide than to superoxide relative to other organisms studied to date. Furthermore, while CAL1 is highly sensitive to aerobiosis, it demonstrates a modest tolerance to exogenous ROS, equivalent to that of the parental organism. It is likely that nonspecific superoxide-protective elements exist in this organism and in some way account for these apparently selective protective qualities.

For many years, the CuZnSOD has been described as being present only in eukaryotes. However, at least 12 prokaryotic CuZnSOD gene sequences (designated sodC) have now been identified (10, 28, 30, 48, 57). Most of these organisms are animal pathogens, and many have host tissue invasion capabilities. CuZnSOD appears to be active in the periplasmic space, as opposed to the cytosol-restricted FeMnSOD, and it has been suggested that it may confer protection against extracellular superoxides such as those produced by activated neutrophils and macrophages. Preliminary attempts to identify the sodC gene in P. gingivalis by PCR methods or to identify its protein product by use of activity gels have been unsuccessful in our laboratory, and a BLAST search (1) of the unfinished genome failed to identify a potential homologue. This may further suggest a superoxide-protective role by nonspecific processes not yet identified.

In conclusion, we have found that SOD expression is dependent upon specific environmental conditions, including oxygen, increased temperature, and alkaline pH. It should be noted that these results are dependent upon the potential limitations of a translational fusion reporter system. However, in this study, results appeared to correspond well with Northern blot analyses of the wild-type organism under similar culture conditions. We have also confirmed that SOD indeed is crucial to the survival of P. gingivalis in aerobic environments as well as under conditions created by specific ROS. However, this enzyme does not appear to protect the organism against human neutrophils and exogenously generated superoxide. By conferring aerotolerance, this enzyme may be essential for colonization and infection in the oral cavity.